diff options
| author | Michael Klishin <mklishin@pivotal.io> | 2017-07-26 19:58:28 +0300 |
|---|---|---|
| committer | Michael Klishin <mklishin@pivotal.io> | 2017-07-26 19:58:28 +0300 |
| commit | 4c9d7492fe31b6e96708661aea51f06183419cab (patch) | |
| tree | ce7798819962dd803c504e3ff2a89794b441881a /docs | |
| parent | fac9eaedf42d10c634fb17b6a100e0e3ad50ddd6 (diff) | |
| download | rabbitmq-server-git-4c9d7492fe31b6e96708661aea51f06183419cab.tar.gz | |
Improve docs
Diffstat (limited to 'docs')
| -rw-r--r-- | docs/rabbitmq.config.example | 12 |
1 files changed, 11 insertions, 1 deletions
diff --git a/docs/rabbitmq.config.example b/docs/rabbitmq.config.example index 7acd6a4d5f..141e737520 100644 --- a/docs/rabbitmq.config.example +++ b/docs/rabbitmq.config.example @@ -236,8 +236,18 @@ %% Fraction of the high watermark limit at which queues start to %% page message out to disc in order to free up memory. + %% For example, when vm_memory_high_watermark is set to 0.4 and this value is set to 0.5, + %% paging can begin as early as when 20% of total available RAM is used by the node. %% - %% Values greater than 0.9 can be dangerous and should be used carefully. + %% Values greater than 1.0 can be dangerous and should be used carefully. + %% + %% One alternative to this is to use durable queues and publish messages + %% as persistent (delivery mode = 2). With this combination queues will + %% move messages to disk much more rapidly. + %% + %% Another alternative is to configure queues to page all messages (both + %% persistent and transient) to disk as quickly + %% as possible, see http://www.rabbitmq.com/lazy-queues.html. %% %% {vm_memory_high_watermark_paging_ratio, 0.5}, |